what is a complex system?

If you ask 10 different people what a complex system is you are likely to get 13 different answers because the term complex system has been used in many disciplines, but not in a consistent way. There is a common theme that brings all these differing uses together though and that theme is an attempt to understand the implications and context of how various types of systems work. In contrast to the classical reductionist approach, which break things down into discrete components for analysis, complexity theory takes a much broader, macro-level view. Complex systems analysis acknowledges that systems (whether in a computer or among the people you work with) do not exist in a vacuum and are influenced by that environment, often in ways that are not seemingly apparent.

Some features of complex systems are that they:

• remember in the sense that what has already happened usually influences what will happen next; however, that influence may be completely unpredictable,


• "learn", meaning that feedback into the system affects the way individual elements (and the system overall, of course) subsequently work. This feedback, therefore, alters elements and system overall, although not necessarily affecting everything in the same way,


• do not have easily defined boundaries as in most complex systems it is not clear where one system begins and another ends,


• can be recursive in the sense that the components of a complex system may be complex systems that themselves contain complex systems and so forth,


• exhibit emergent behavior, which are properties of the interactions of the system that can only be studied at a high level. In most cases, these behaviors are undetectable at the element level.

In future posts, my intent is to take a look at how these concepts apply to libraries as well as begin to explore the implications of this. A nice overview of complex systems by Gershenson and Heylighen can be found at http://uk.arxiv.org/ftp/nlin/papers/0402/0402023.pdf